Messages per Month
| Subject: | Re: JNI intBitsToFloat return error | |
|---|---|---|
| From: | phon...@mobileandembedded.org (phon...@mobileandembedded.org) | |
| Date: | Mar 27, 2007 7:14:27 am | |
| List: | net.java.dev.phoneme.advanced | |
I get pretty much the same thing on x86:
arg v = 1120442778 (0x42c8999a) u.i = 1120442778 (0x42c8999a) u.f = 1073741824 (0x40591333) 100.300003 (jfloat)u.f = 100.300003 *(jfloat *)&u.i = 100.300003
The difference is in the first two values in the 3rd line. However, I think this
may be the result of a bug in the printf call. You are passing in a float type,
but using %d and %x to display the value. Due to calling convention differences
between floats and ints, I don't think this is supported.
Chris [Message sent by forum member 'cjplummer' (cjplummer)]





